One-step photodeposition of spatially separated CuOx and MnOx dual cocatalysts on g-C3N4 for enhanced CO2 photoreduction

Photocatalytic CO 2 conversion into valuable chemicals has been proved to be a promising strategy for relieving energy shortage and environmental pollution. Nevertheless, the rapid recombination of photogenerated carriers of photocatalyst greatly limits their actual application. In this work, dual CuO x and MnO x cocatalysts are decorated on g-C 3 N 4 nanosheets via a one-step photodeposition strategy. Bene�ting from the repulsion between Cu 2+ and Mn 2+ cations, a novel g-C 3 N 4 -based heterostructure loaded with spatially separated CuO x nanoparticles and MnO x nanosheets dual cocatalysts has been successfully fabricated. The Cu favors the trapping of electrons, while MnO x tends to collect holes. Moreover, the Cu 2 O/g-C 3 N 4 p-n heterojunction also accelerates the charge separation. As a result, the photogenerated holes and electrons �ow into and out of the photocatalyst, respectively, resulting in enhanced charge separation for achieving e�cient CO 2 photoreduction over CuO x /g-C 3 N 4 /MnO x . Impressively, the optimized CuO x /g-C 3 N 4 /MnO x exhibits an improved CO production rate of 5.49 µmol g − 1 h − 1 , which exceeds over 27.5 times than bare g-C 3 N 4 . This work designs a promising photocatalyst for CO 2 photoreduction and develops a novel one-step photodeposition route for decorating spatially separated dual cocatalysts on a photocatalyst.


Introduction
Converting CO 2 into useful chemicals is an e cient strategy to solve greenhouse effect (Kondratenko et  In recently years, graphitic carbon nitride (g-C 3 N 4 ) has achieved increasing interests in CO 2 photoreduction because of its low cost and visible-light response features and sutiable band structure 2020b). A cocatalyst is commonly considered to be an intrinsically inactive material which can enhance the activity, stability and selectivity of single photocatalyst. Generally speaking, a cocatalyst play two important roles in improving the photocatalytic e ciency: i) capturing the photogenerated carriers for restraining the recombination of carriers (Li et al. 2019b); ii) offering more active sites for boosting photocatalytic reactions (Yang et al. 2020  ). Numerous works have been highly done to load appropriate dual cocatalysts on the surface of the host photocatalyst in unifrom distribution. Noteworthily, dual cocatalysts deposited by the two-step method would result in an overlapped of oxidation and reduction cocatalysts, which is not expected in photocatalytic performance improving. Unfortunately, this issue has been seldom solved in report, and it is worthy to further study relevant photocatalytic mechanism.
In general, a repulsive force is always existed between two metal ions with the same positive charges (Fenton et al. 2019). Therefore, they tend to attach to different adsorption sites on the surface of the substrate (Fenton et al. 2018). With the obove considerations in mind, the spatial separation of oxidation and reduction cocatalysts can be achieved by the one-step photodeposition using the repulsion between two metal cations. In this work, Cu 2+ and Mn 2+ were used as the metal precursors to construct the CuO x /g-C 3 N 4 /MnO x composite structure in which CuO x and MnO x are loaded separately on g-C 3 N 4 surface by one-step photo-reduction route (Fig. 1) Brie y, a xed amount of dried urea was put in a crucible with a lid for keeping at 550 ℃ for 4 h with a heating rate of 15 ℃ min − 1 in Mu e oven. After being cooled down to 25 ℃, the uffy porous g-C 3 N 4 powder was obtained.
Synthesis of CuO x /g-C 3 N 4 /MnO x composites The CuO x /g-C 3 N 4 /MnO x was prepared via a facile photoreduction approach. Typically, 200 mg g-C 3 N 4 was added in 100 mL distilled water with assistance of sonication to get well-dispersed homogenous suspensions. After that, 6.98 mg CuSO 4 •5H 2 O and 4.77 mg MnSO 4 •H 2 O were quickly added. The mixtrue was further stirred for 4 h under the irradiation of 300 W Xe lamp. After ltration and washing with distilled water several times, the obtained CuO x /g-C 3 N 4 /MnO x hybrids was dried at 100°C under vacuum for 10 h. The prepared products are named as CuO x /g-C 3 N 4 /MnO x -x, where "x" is 0.5, 0.75, 1.0, 3.0 and 5.0, which expresses the weight ratios of both CuO x and MnO x to g-C 3 N 4 . The samples with 1.0 wt% weight ratio of CuO x to g-C 3 N 4 and 1.0 wt% weight ratio of MnO x to g-C 3 N 4 were also prepared, and labled as CuO x /g-C 3 N 4 -1 and MnO x /g-C 3 N 4 -1, respectively.

Synthesis of CuO x /(MnO x /g-C 3 N 4 )-1 and MnO x /(CuO x /g-C 3 N 4 )-1 composites
A two-step photoreduction method was used to synthesis CuO x /(MnO x /g-C 3 N 4 )-1 and MnO x /(CuO x /g-C 3 N 4 ). Typically, 200 mg g-C 3 N 4 was added in 100 mL distilled water with assistance of sonication to get well-dispersed homogenous suspensions. And then a certain amount of CuSO 4 •5H 2 O was added. The mixtrue was further stirred for 4 h under the irradiation of 300 W Xe lamp. After ltration and washing with distilled water several times, the CuO x /g-C 3 N 4 hybrid can be obtained after dring at 100°C under vacuum for 10 h. Then, the obtained CuO x /g-C 3 N 4 was dispersed in 100 mL distilled water and subsequently a certain amount of MnSO 4 •H 2 O was added into the solution. After the same 4 h photoreduction and drying processes, the CuO x /(MnO x /g-C 3 N 4 )-1 was obtained. Similarly, MnO x /(CuO x /g-C 3 N 4 )-1 sapmle was prepared using the above procedure by switching the addition of CuSO 4 •5H 2 O and

Results And Discussion
XRD measurements were conducted to characterize the crystal structures of the obtained samples. As displayed in Fig. 2a, the XRD pattern of g-C 3 N 4 displayed a weak peak at 13.1°, attributing to (100) peak originated from the in-plane ordering of tri-s-triazine units. The strong diffraction peak at 27.4° was assigned to the dense interlayer-stacking (002) peak of aromatic segment (Tian et al. 2018). Notably, no peaks of CuO x or MnO x were detected in the XRD patterns of CuO x /g-C 3 N 4 /MnO x composites, attributing to the low content, high dispersion and amorphous structures of CuO x and MnO x (Xiang et al. 2021).
Moreover, the diffraction peaks also displayed no obviously changes after the modi cation with CuO x and MnO x , indicating that g-C 3 N 4 maintained a good stability during the photodeposition reaction.
Scanning electron microscope (SEM) and transmission electron microscope (TEM) measurements were utilized for studying the morphology and structural feature of the obtained samples. As the SEM image shown in Fig. 2b and 2c, it is clear that both g-C 3 N 4 and CuO x /g-C 3 N 4 /MnO x -1 exhibit graphite-like layered structure with curled and wrinkled surfaces, indicating that the g-C 3 N 4 morphology was almost unchanged after being modi ed with cocatalysts. High-resolution TEM (HRTEM) were conducted to investigate the distribution of CuO x and MnO x . It can be seen that the spatially separated CuO x nanoparticles and MnO x nanosheets were successfully deposited on the bare g-C 3 N 4 layers ( Fig. 2d and   e). In addition, no fringes belong to CuO x or MnO x are observed in the HRTEM image because of their amorphous structures, which was in accordance with the XRD results. Moreover, the related elemental mapping analysis also displays the homogeneous dispersion of Cu, Mn and O elements, indicating the high dispersions of CuO x nanoparticles and MnO x nanosheets on the g-C 3 N 4 surfaces (Fig. 2f).
XPS measurements were further conducted to investigate the surface compositions of the obtained samples. The high-resolution C 1s spectra of g-C 3 N 4 can be divided into three peaks at 284.5, 285.3 and 287.8 eV (Fig. 3), attributing to the the adventitious carbon, sp 3 -bonded carbon species from defects and  In which α, h, ν, A, and E g are the absorption coe cient, Planck constant, light frequency, a constant, and band gap energy, respectively. Accordingly, the band gap can be measured from the intercept of the tangent of the curve of (αhν) 1/2 against the radiant energy hν. As shown in Fig. 3e, the band gap of g-C 3 N 4 is calculated to be 2.64 V and the band gaps of CuO x /g-C 3 N 4 /MnO x -1, CuO x /g-C 3 N 4 /MnO x -3 and CuO x /g-C 3 N 4 /MnO x -5 composites are 2.53, 2.47 and 2.36, respectively (Fig. 3e). These results com rm the improved light harvesting e ciency in the visible light region.
Generally, the higher photocurrent density implies more e cient photogenerated electron-hole pairs separation and transportion across interface (Wang et al. 2018b). Therefore, the improved charge separation of the obtained g-C 3 N 4 -based composites can be further veri ed by photocurrent measurements under visible irradiation. As shown in Fig. 3f, the CuO x /g-C 3 N 4 /MnO x -1 composite shows the highest photocurrent density than bare g-C 3 N 4 , CuO x /g-C 3 N 4 -1 and MnO x /g-C 3 N 4 -1, implying that photogenerated charge carrier separation e ciency is signi cantly enhanced by dual cocatalysts modi cation.
The photocatalytic CO 2 reduction activities of the obtained nanocomposites were evaluated under visible light illumination (> 420 nm) in the gas-solid system. As shown in Fig. 4a, g-C 3 N 4 nanosheets exhibited relatively lower CO evolution with a rate of 0.2 µmol g -1 h -1 , resulting from its rapid charge recombination.
Notably, the CO productions were greatly improved with the increased loading of CuO x and MnO x cocatalysts, and the production of CO achieved the highest rate (5.49 µmol g -1 h -1 ) over CuO x /g-C 3 N 4 /MnO x -1, exceeding approximately 27.5 times that of bare g-C 3 N 4 . Additionally, the apparent quantum e ciency of CuO x /g-C 3 N 4 /MnO x -1 at 420 nm is calculated to be 0.9%. While further increasing CuO x and MnO x amounts would be detrimental to the photocatalytic performance, attributing to the fact that the overloading of CuO x and MnO x could obscure active sites. Recycled of CuO x /g-C 3 N 4 /MnO x -1 still maintained approximately 81% of its initial activity after four cycling runs (each run for 4 h), indicating a good stability (Fig. 4c).
To unveil the electron transfer path and the mechanism of the CuO x and MnO x as dual cocatalysts for photocatalytic CO 2 reduction, a series of comparative experiments have been conducted. As shown in Fig. 4b, the CO yield of ternary CuO x /g-C 3 N 4 /MnO x -1 photocatalyst was even higher than the sum of that of MnO x /g-C 3 N 4 -1 and CuO x /g-C 3 N 4 -1. This result can be attributed to the dual cocatalysts possessing a synergistic effect on suppressing the photogenerated charges recombination and furnishing the photocatalytic reaction with more active sites. Notably, the CO production rates of the catalysts with CuO x or MnO x follow the same orders of MnO x /g-C 3 N 4 -1 CuO x /g-C 3 N 4 -1, indicating that the oxidation cocatalyst MnO x has more to do with the CO 2 reduction than the reduction cocatalyst CuO x . This result could be attributed to the fact that MnO x can collect photoinduced holes to accelerate the H 2 O oxidiation reactions, thus promoting the photocatalytic CO 2 reduction process. Additionally, the photocatalytic activity of CuO x /g-C 3 N 4 /MnO x -1 was also higher than MnO x /(CuO x /g-C 3 N 4 ) and CuO x /(MnO x /g-C 3 N 4 ), demonstrating the spatially separated dual cocatalysts are more conducive to improving the separation e ciency of photoinduced carriers.
In-situ FTIR spectroscopy was performed to explore the photocatalytic mechanism by investigating the key reaction intermediate products of the reduction of CO 2 and H 2 O toward CO. The spectra were recorded every 5 min. As presented in Fig. 5a, with the illumination time from 0 to 30 min, the intensities of some new peaks that appeared in the in-situ FTIR results of CuO x /g-C 3 N 4 /MnO x -1 gradually increased.
The absorption peaks at 1283 and 1440 cm -1 are assigned to bicarbonate ( indicating that the CO* is easily desorbed on the CuO x /g-C 3 N 4 /MnO x -1 surface and turns into the anl CO product. According to the in-situ FTIR results, the possible reaction path for CO 2 photoreduction over CuO x /g-C 3 N 4 /MnO x composites is proposed as follows (the asterisks denote catalytically active sites): 1 , or In order to investigate the photogenerated charges transfer behavior during the photcatalytic process, Mott-Schottky plots were performed for investigating the in uence of CuO x and MnO x cocatalyst on band structure g-C 3 N 4 , as well as the semiconductor types. As shown in Fig. 5b, the positive curve slope unveils that the introduction of CuO x and MnO x did not change the n-type properties of g-C 3 N 4 (Tian et al. 2018).
Moreover, the at band potential can be determined via the Mott-Schottky equation : In which C, e, ε, ε 0 and N represent the capacitance of the space charge region, electron charge, dielectric constant, vacuum permittivity and the carrier density of samples, respectively. E represents the electrode applied potential and E fb is the at band potential. T and k represent the absolute temperature and Boltzmann constant, respectively. The kT/e term can be neglected at room temperature because of the extremely small vaule (25.693 meV). Accordingly, the E fb of g-C 3 N 4 is measured to be -0.90 V (vs. SCE).
The CB potential of g-C 3 N 4 is estimated to be -0. Given the above results, a probable mechanism of CuO x /g-C 3 N 4 /MnO x composite for CO 2 photoreduction to CO has been proposed. As shown in Fig. 5c, when CuO x nanoparticles are attached on g-C 3 N 4 , a

Conclusions
In summary, we have prepared various CuO x /g-C 3 N 4 /MnO x heterostructures through a simple one-step photodeposition approach. When CuO x /g-C 3 N 4 /MnO x was illuminated by light, the photogenerated electrons were migrated from the CB of g-C 3 N 4 to Cu species for the reduction of CO 2 , and the related holes were captured via MnO x nanoparticles to oxidize water. Besides, the Cu 2 O/g-C 3 N 4 p-n heterojunction also accelerates the charge separation. As an outcome, the spatially separated CuO x and MnO x dual cocatalysts greatly restrain the recombination rate of photogenerated carriers and facilitate the CO 2 conversion kinetically, resulting in the enhanced CO 2 -to-CO photoreduction performance. This work not only provides a high-performance photocatalyst for CO 2 reduction, but also develops a novel strategy for decorating spatially separated dual cocatalysts on a photocatalyst.

Declaration of competing interest
The authors declare no competing nancial interests. Figure 1 Schematic diagram of the superiority of one-step photoreduction method.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.